Alzheimer's disease (AD) looms as the major public health problem of the twenty-first century. A variety of risk factors for developing AD have been proposed including a strong genetic component. Even though familial clustering of AD is not common, it is unclear whether genetic mechanisms account for AD in only a small proportion of cases or whether, in fact, all AD is inherited. The overall goals of this proposal are (1) to determine what proportion of AD cases are very likely sporadic and familial cases of AD for subsequent studies of clinical profiles which may lead to a better understanding of the etiology of the disease. In order to accomplish these goals we shall establish a family registry using data collected by researchers in 6 centers in the United States and Canada. The diagnosis of AD will be established by NINCDS/ADRDA research criteria. Family history and clinical history data will be elicited from consecutively ascertained patients with suspected dementia by direct and telephone interview using standardized questionnaire instruments. Medical records, autopsy reports and death certificates will be sought and carefully reviewed by trained personnel. These data will be assembled in a computerized data base. We expect to collect information on 700 examined cases (50%-65% AD, 35%-50% non-AD dementia) per year totaling 3500 cases by the end of the project period. Sex and age adjusted rates of dementia in the general population will be derived from 3 control groups: (1) the Framingham Study cohort comprised of 5,209 individuals; (2) the control group for the University of Washington Alzheimer Disease Patient Registry, the cases of which will also be contributed to this study; (3) a stratified sample of 4,164 community residents participating in an epidemiological study of the elderly at Duke University. These data will allow us to estimate the age at onset distribution of AD using complex life table analysis techniques, to evaluate genetic and environmental transmission models for AD by pedigree analysis, and to develop a strategy for estimating the likelihood that a case is non- familial. Previous studies addressed these questions less effectively because of much smaller sample sizes. Identification of variant transmission patterns among geographic areas or among persons with particular risk factors may yield important clues about etiology. Progress in these areas will hopefully lead to better methods for identifying genes involved in expression of AD and for assessing risk of AD in individuals according to family history and other risk factors.